𝔖 Bobbio Scriptorium
✦   LIBER   ✦

The computational brain: Patricia S. Churchland and Terrence J. Sejnowski (MIT Press, Cambridge, MA, 1992); xi, 544 pages, $39.95

✍ Scribed by George N. Reeke Jr


Book ID
102989476
Publisher
Elsevier Science
Year
1996
Tongue
English
Weight
911 KB
Volume
82
Category
Article
ISSN
0004-3702

No coin nor oath required. For personal study only.

✦ Synopsis


The Computational Bruin represents a notable attempt to summarize the current status of research on computational models of brain function at an introductory level. As we might expect from the authors' backgrounds, the work integrates topics in philosophy, neuroscience, and connectionist modelling, and it will therefore be reviewed in all three areas after an outline of the contents.

The book begins with a background chapter on neuroscience. Following Shepherd [ 3 11, the authors lay out seven levels of description (behavior, systems and pathways, local circuits, neurons, microcircuits, synapses, membranes and molecules) that must be considered to understand the brain. The question, "What are the most basic structural features relevant to neural computation?' is asked, and answered with a list of likely "must know" items that includes such things as specialization of function in regions of the nervous system, numbers and connectivity of neurons and synapses, spiking, timing of neural signals, receptive fields, and parallel architecture. This is followed by a description of some computational organizations that have been proposed as models for neural systems. These include linear associators, Hopfield networks, Boltzmann machines, competitive learning and back-propagation, feedforward and recurrent nets. The authors close this section with a strong statement that the method of parameter adjustment in a model network is not important so long as the model network and the real network "end up at locations on the error surface that are rather close" (p. 131). This assertion permeates the authors' entire approach to the subject in a way that I find most troubling, and I shall return to it at some length below.

Churchland and Sejnowski go on to explore the problem of representations, taking examples mainly from vision to describe concepts such as vector coding and state * (MIT Press,


πŸ“œ SIMILAR VOLUMES


The mind-brain continuum: Sensory proces
πŸ“‚ Article πŸ“… 1997 πŸ› Elsevier Science 🌐 English βš– 106 KB

Case-Based ReasoninQ: Experiences. Lessons. ~ Future Directions. Edited by David B. Leake. AAAI Press/ MIT Press, Menlo Park, CA/Cambridge, MA. (1996). 420 pages. $40.00. ## Contents: Contributors. Preface. 1. CBR in context: The present and future (David B. Leake). 2. A tutorial introduction to